Paper from CHI 2008 Proceedings - Improved Video Navigation and Capture. It present a method for browsing videos by directly dragging their content, in contrast with traditional seeker bar which focus on visual content rather than time.

1. Video Browsing by Direct
Manipulation
Pierre Dragicevic, Gonzato Ramos, Jacobo Bibliowicz,
Derek Nowrouzezahrai, Ravin Balakrishman,
Karan Singh
User Interface Design 646
Presented by Vashira Ravipanich
5171439021

2. Introduction
• All video players use
“seeker bar” to control
user interaction
• What if you can directly
dragging in the movie?

3. Introduction
• This paper presents a method for browsing
videos by “directly dragging” their content
• Automatically extracting motion data
• Relative Flow Dragging

5. Why Direct Manipulation?
• Input ~ Output
• Time V.S. Space
• Both are complementary NOT rival
Input like finger move = Output like
mouse movement
Time = seeker Bar, Space = Direct
Manipulation

6. How does it works?
• Videos = sequence of multiple pictures
(frame)
• Extract object(s) movement Call “Trajectory
Extraction”
• Construct “hint path”

7. Relative Flow Dragging
• Directness Directness => user input lang ==
generated output
• Matching gesture with motion
2D = map
3D = scaling object, rotating object

8. Type of dragging
• Curvilinear Dragging
• Flow Dragging
• Relative Dragging

9. Direct Manipulation Video Player - DIMP

10. Background Stabilization

11. Position Feedback

12. Trajectory Extraction
• Computer Vision Approaches
• Object Tracking
- object on video sequence
- motion capture, surveillance
• Optical Flow
- whole picture, calculate pixels
- video compression
• Optical Flow is better for general video player

13. Curvilinear Dragging Design

14. Proposed Solutions
• 3D Distance Method
• (x, y, z) where z is arc-length distance from
the curve origin

15. Limitations
• Video with back-and-forth movement, i.e a
couple dancing tango
• DIfficult to visualize path clearly

16. Evaluation
• User Study
• 6 males, 10 females
• 18 - 44 years old
• Test with 2 videos with given objectives
• Offer both seeker bar and relative flow
dragging
• Which one user comfortable with the most?

18. Quantitative Results

19. Quantitative Results

20. Previous work on Video Browsing
• Non-Linear Video Browsing
- Segment of difference importance
- Estimating motion activity
• Visual Summaries
- Generate mosaic from key frames
• Content-Based Video Retrieval

21. Conclusion & Future Work
• New way of browsing videos using direct
manipulation
• Appealing to touch-input handheld. iPhone,
Pocket PC.
• Interactive Learning Environments.

22. References
1. Accot, J. and Zhai, S. (1997). Beyond Fitts' law: mod- 11. Dragicevic, P., Huot, S. and Huot, S. (2002). SpiraC-
els for trajectory-based HCI tasks. CHI. p. 295-302. lock: a continuous and non-intrusive display for up-
2. Appert, C. and Fekete, J. (2006). OrthoZoom scroller: coming events. CHI Extended Abstracts. p. 604-605.
1D Multi-Scale Navigation. CHI. P. 21-30. 12. Goldman, D.B., Curless, B., Salesin, D. and Seitz, S.M.
3. Autodesk Maya. http://www.autodesk.com/ (2006). Schematic storyboarding for video visualization
4. Baudel, T., Fitzmaurice, G., Buxton, W., Kurtenbach, and editing. SIGGRAPH. p. 862-871.
G., Tappen, C. and Liepa, P. (2002). Drawing system 13. Guimbretière, F. (2000). FlowMenu: combining com-
using design guides. US Patent # 6,377,240. mand, text, and data entry. UIST. p. 213-216.
5. Beauchemin, S.S. and Barron, J.L. (1995). The compu- 14. Hölzl, R. (1996). How does ‘dragging’ affect the learn-
tation of optical flow. ACM Computing Surveys, 27(3). ing of geometry? International Journal of Computers
p. 433-467. for Mathematical Learning, 1(2). p. 169-187.
6. Beaudouin-Lafon, M. (2000). Instrumental Interaction: 15. Hutchins, E.L., Hollan, J.D. and Norman, D.A. (1987).
An interaction model for designing post-WIMP user in- Direct manipulation interfaces. In Human-Computer in-
terfaces. CHI. p. 446-453. teraction: A Multidisciplinary Approach. R. M. Baeck-
7. Beaudouin-Lafon, M. (2001). Novel interaction tech- er, Ed. Morgan Kaufmann. p. 468-470.
niques for overlapping windows. UIST. p. 153-154. 16. Irani, M., Anadan, P. and Hsu, H. (1995). Mosaic based
8. Bezerianos, A., Dragicevic, P. and Balakrishnan, R. representations of video sequences and their applica-
(2006). Mnemonic rendering: an image-based approach tions. Intl. Conference on Computer Vision. p. 605-611.
for exposing hidden changes in dynamic displays. 17. Kim, C. and Hwang, J. (2002). Fast and automatic
UIST. p. 159-168. video object segmentation and tracking for content-
9. Buxton, W. (1986). There's more to interaction than based applications. IEEE Trans. Circuits and Systems
meets the eye: some issues in manual input. In User for Video Technology, 12. p. 122-129.
Centered System Design: New Perspectives on Human- 18. Kimber D., Dunnigan, T., Girgensohn, A., Shipman, F.,
Computer Interaction. Lawrence Erlbaum. p. 19-337. Turner, T. and Yang, T. (2007). Trailblazing: Video
10. Cheng,Y. (1995). Mean shift, mode seeking, and clus- playback control by direct object manipulation. ICME.
tering. IEEE Transactions on Pattern Analysis and Ma- p. 1015-1018.
chine Intelligence, 17(8). p. 790-799. 19. Li, F.C., Gupta, A., Sanocki, E., He, L. and Rui, Y.

23. Thank you